Friday, June 20, 2008

Total Value - YTD leaders

If you believe in voting for your all stars based on the first 2.5 months of performance, then here's a list for you. It's the current top-ranking players at each position, based on my total value rankings, the methodology for which is described here in detail. Briefly, though:

RAR refers to Runs Above Replacement, an offense-only assessment of value similar to VORP without position adjustments. It is based on linear weights.

FLD refers to +-fielding runs, and is calculated by translating Zone Rating (from STATS Inc) and Revised Zone Rating (from Baseball Info Solutions) into +-runs statistics, and then taking the average of those two fielding estimates. Read it as runs saved vs. an average defender at the same position.

TV is Total Value, a measure of total value (in runs) above a replacement player. It is RAR + fielding + a position adjustment (not shown) to put all positions on the same level. Note that I am imposing a -20 runs/700 PA floor on [fielding+position adjustment], because no one should be less valuable (in terms of intrinsic value, not impact on a team, of course) on the field on a per-pa basis than a DH.

Click here to access YTD total value estimates for all MLB players. Data are through June 15th. I'll post updates periodically.

Catchers
NL
Brian McCann (ATL): 23 RAR, +3 FLD, 29 TV
Geovany Soto (CHN): 19 RAR, +1 FLD, 24 TV

AL
Joe Mauer (MIN): 17 RAR, +3 FLD, 23 TV
Dioner Navarro (TB): 9 RAR, +3 FLD, 15 TV

First Basemen
NL
Lance Berkman (HOU): 46 RAR, +9 FLD, 51 TV
Albert Pujols (STL): 39 RAR, +9 FLD, 45 TV

AL
Kevin Youkilis (BOS): 22 RAR, +3 FLD, 22 TV
Casey Kotchman (LAA): 13 RAR, +8 FLD, 17 TV

Second Basemen
NL
Chase Utley (PHI): 36 RAR, +7 FLD, 43 TV
Dan Uggla (FLA): 29 RAR, +1 FLD, 30 TV

AL
Mark Ellis (OAK): 11 RAR, +11 FLD, 22 TV
Brian Roberts (BAL): 22 RAR, -1 FLD, 21 TV

Third Basemen
NL
Chipper Jones (ATL): 43 RAR, +11 FLD, 54 TV - NL MVP thus far, by my numbers
Troy Glaus (STL): 15 RAR, +8 FLD, 23 TV

AL
Alex Rodriguez (NYY): 26 RAR, +4 FLD, 30 TV (in 214 PA's!)
Scott Rolen (TOR): 11 RAR, +13 FLD, 24 TV

Shortstops
NL
Hanley Ramirez (FLA): 26 RAR, -0 FLD, 28 TV
Rafael Furcal (LAN): 22 RAR, +1 FLD, 23 TV

AL
Orlando Cabrera (CHA): 1 RAR, +8 FLD, 11 TV - What's up with AL shortstops?
Michael Young (TEX): 12 RAR, -5 FLD, 9 TV

Left Field
NL
Pat Burrell (PHI): 31 RAR, +4 FLD, 33 TV
Jason Bay (PIT): 29 RAR, -1 FLD, 26 TV

AL
Johnny Damon (NYY): 26 RAR, +2 FLD, 27 TV
Jacoby Ellsbury (BOS): 18 RAR, -1 FLD, 17 TV

Center Field
NL
Carlos Beltran (NYN): 18 RAR, +7 FLD, 27 TV
Nate McLouth (PIT): 30 RAR, -12 FLD, 21 TV

AL
Grady Sizemore (CLE): 31 RAR, +3 FLD, 35 TV - AL MVP thus far, by my numbers. Leadoff PA's help.
B.J. Upton (TB): 23 RAR, +2 FLD, 26 TV

Right Field
NL
Brian Giles (SDN): 25 RAR, +12 FLD, 34 TV
Ryan Ludwick (STL): 28 RAR, -6 FLD, 22 TV

AL
J.D. Drew (BOS): 25 RAR, +3 FLD, 25 TV
Nick Markakis (BAL): 21 RAR, +1 FLD, 20 TV

DH (defined here as the top remaining unselected hitters):
NL
Adrian Gonzalez (SDN): 26 RAR, +3 FLD, 26 TV
Aramis Ramirez (CHN): 23 RAR, -4 FLD, 20 TV

AL
Milton Bradley (TEX): 34 RAR, -5 FLD, 26 TV
Josh Hamilton (TEX): 27 RAR, -4 FLD, 23 TV


Most Valuable Position Players Overall

1. Chipper Jones (ATL): 54 TV
2. Lance Berkman (HOU): 51 TV
3. Albert Pujols (STL): 45 TV
4. Chase Utley (PHI): 43 TV
5. Grady Sizemore (CLE): 35 TV
6. Brian Giles (SDN): 34 TV
7. Pat Burrell (PHI): 33 TV
8. Dan Uggla (FLA): 30 TV
9. Alex Rodriguez (NYY): 30 TV (in 214 PA's!)
10. Brian McCann (ATL): 29 TV

I find it pretty interesting how overwhelmingly the NL players are leading in total value. Part of that is that is probably a methods problem: even though I use different replacement levels for the two leagues, I'm using the same MLB average runs per game for both leagues. Right now, though, AL hitters are hitting more poorly than NL hitters. But even that difference can't be worth more than a couple of runs at this point in the season.

On to the pitchers!

NL Starters (I'm using FIP Runs to try to get around variation in defense and luck)
1. Brandon Webb (ARI): 29 RAR
2. Tim Lincecum (SFN): 27 RAR
3. Edinson Volquez (CIN): 25 RAR
4. Cole Hamels (PHI): 21 RAR
5. Dan Haren (ARI): 21 RAR
6. Aaron Harang (CIN): 20 RAR

NL Relievers
1. Kerry Wood (CHN): 17 RAR
2. Brad Lidge (PHI): 16 RAR
3. Jon Rauch (WAS): 13 RAR
4. Takashi Saito (LAN): 12 RAR

AL Starters
1. Roy Holliday (TOR): 32 RAR
2. Cliff Lee (CLE): 29 RAR
3. Jose Contreras (CHA): 25 RAR
4. Javier Vazquez (CHA): 24 RAR
5. C. C. Sabathia (CLE): 23 RAR
6. Ervin Santana (LAA): 23 RAR

AL Relievers
1. Mariano Rivera (NYY): 16 RAR
2. Jonathan Papelbon (BOS): 14 RAR
3. Ramon Ramirez (KC): 10 RAR
4. Bobby Jenks (CHA): 9 RAR

AL pitching looks stronger than NL pitching by these measures, though it's not overwhelming. FWIW, while I haven't checked, I do expect Harang's poor performance today to knock him off these charts.

If the above list was used to pick all-star teams, with 28 players per team, here's how it would break down in terms of representation by team (assuming no minimum representation requirement):
ARI 2
ATL 2
BAL 2
BOS 4
CHA 4
CHN 3
CIN 2
CLE 3
COL --
DET --
FLA 2
HOU 1
KC 1
LAN 2
LAA 2
MIL --
MIN 1
NYA 3
NYN 1
OAK 1
PHI 4
PIT 2
SD 2
SEA --
SF 1
STL 3
TB 2
TEX 3
TOR 2
WAS 1
If anyone out there is interested in the raw spreadsheets (and they are a bit "raw," because I'm notoriously messy with my Excelling) that I use to calculate these data, feel free to drop me an e-mail and I'll send them your way. They're big, though, so I will ask that you be able to extract 7-zip documents. But that's what everyone should be using anyway! :)

Update on Outfielders:
If you prefer to lump all of your outfielders together (which, of course, is how the all-star voting works and is probably how I should have done it anyway), they come out looking like this:

NL
1. Brian Giles (SDN): 25 RAR, +12 FLD, 34 TV
2. Pat Burrell (PHI): 31 RAR, +4 FLD, 33 TV
3. Carlos Beltran (NYM): 18 RAR, +7 FLD, 27 TV
4. Jason Bay (PIT): 29 RAR, -1 FLD, 26 TV
5. Nate McLouth (PIT): 30 RAR, -12 FLD, 21 TV
6. Ryan Ludwick (STL): 28 RAR, -6 FLD, 22 TV

AL
1. Grady Sizemore (CLE): 31 RAR, +3 FLD, 35 TV
2. Johnny Damon (NYY): 26 RAR, +2 FLD, 27 TV
3. B.J. Upton (TB): 27 RAR, +2 FLD, 26 TV
4. J.D. Drew (BOS): 25 RAR, +3 FLD, 25 TV
5. Josh Hamilton (TEX): 27 RAR, -4 FLD, 23 TV
6. Nick Markakis (BAL): 21 RAR, +1 FLD, 20 TV

No differences for the NL, but John Hamilton makes the cut instead of Jacoby Ellsbury in the AL. Jason Giambi (24 RAR, -6 FLD, 17 TV) would be Josh Hamilton's replacement at DH on the AL team.

8 comments:

  1. Shouldn't Manny, not Ellsbury, be listed in LF? BBRef has him there in 53 games this season.

    ReplyDelete
  2. Manny's also listed as a LF--he just gets dinked down to 15 RAR because of his defense.

    Players are pigeonholed into positions based on where they've played the most. Ellsbury has 35 G in LF and 31 G in CF, which is how he got placed in that position. In this case, though, that's apparently the result of defensive shifts, because Ellsbury has started the most in CF: 28 G in CF vs. 17 G in LF. So, if you like, you can replace Ellsbury on this list with Jack Cust (16 RAR). -j

    ReplyDelete
  3. Justin, nice work. A question about the 0.5 penalty for NL players. Is this for hitters only? If it's only for hitters for example, does this mean that an NL team that finished with 90 wins, would have won 90-(9 * 0.5) 85.5 wins in the AL (in a league with no interleague games)?
    Just wondering.
    vr, Xeifrank

    ReplyDelete
  4. Xeifrank, there is a "penalty" for both hitters and pitchers. The AL's a more challenging league, and therefore an "MLB average player" will do worse there than in the NL. And yet, because they are MLB average players, they should have the same value on the free agent market no matter where they play. 0.5 win (5 run) penalty helps account for that.

    Extrapolating to team wins is an interesting question... ... Honestly, I'd have to think about the exact consequences of how these things would add up. I *think* you're right about the hitters (though using pythagenpat would be a more accurate way to model things), but there would need to be an adjustment for pitchers as well. And I'm having a hard time believing that the difference between leagues is ~9 full wins (assuming equal effect on hitters and pitchers), but maybe that's true. This'd be a good question for Tango. :)
    -j

    ReplyDelete
  5. Justin:

    McClouth being that much of a butcher on defense is a surpise.

    Thanks for this.

    ReplyDelete
  6. Justin, you've obviously done something wrong -- ARod and Youkilis are two of the most overrated players in the majors according to the players themselves. How could they possibly be All-Stars?

    Love that you're posting the numbers for the whole league. Thanks.

    ReplyDelete
  7. @Sky, I think it must be because the rest of the AL sucks. :)

    Anyway, yeah, I've finally got my spreadsheets set up to allow me to post mlb-wide numbers pretty quickly. I'll eventually get this thing automated enough to update automatically, but I'm not quite there yet. :)
    -j

    ReplyDelete
  8. Forget auto-updating, when do we get the tables with filters and one-click sorting? ; )

    Oh, and I'll take a copy of the spreadsheet you're using, thanks.

    ReplyDelete